An analysis framework based on the mismatch between business cycle improvements and market pricing reveals a clear divergence between earnings upgrades and stock price movements in the A-share market. During the week of August 10th to August 14th, sector performance continued to show clear differentiation. Based on weekly changes and turnover data for the Shenwan primary industries, a clear market structure emerges: aside from the comprehensive sector, telecommunications led the gains. This was followed by the "carbon-based sectors" we have frequently recommended, including pharmaceuticals, the real estate chain, and food and beverage, highlighting a pronounced rotation pattern. Conversely, non-ferrous metals led the declines, with sectors like non-bank finance, beauty and personal care, transportation, and basic chemicals also experiencing pullbacks. At the macro level, after three consecutive weeks of gains, the Shanghai Composite Index saw a slight 0.33% decline for the week, indicating a degree of market volatility and divergence. For instance, while the ChiNext index posted positive returns for the week, the STAR 50 index was negative, reflecting market disagreement regarding the momentum for further upside and the main structural theme going forward. We believe the market has already undergone a certain degree of recovery from its 3741-point level. Looking ahead, patience is required; an overly aggressive stance is not advisable. The focus should be on medium-term investment and value investing, concentrating on uncovering structural opportunities.

With limited marginal changes on a weekly basis, we believe that repairing high-quality blue-chip stocks is the core medium-term direction for the A-share market. A "mismatch" in the market typically refers to a situation where the market is slow to react to, or even prices in the opposite direction of, improvements in a particular asset's fundamentals. We have constructed a simple yet powerful framework using consensus forecast data: "magnitude of earnings upgrades minus year-to-date price change." A positive difference indicates that the market has raised its earnings forecasts for a sector in 2026, but the stock price has not kept pace, or has even fallen, representing a quantifiable signal of being "wrongly sold off." Specifically, the "magnitude of earnings upgrades" is calculated as the two-year average of the percentage change in the consensus forecast for net profit attributable to the parent company for the full years 2026 and 2027 compared to the forecast value on January 1, 2026. This smoothes out noise from single-year forecast adjustments and reflects the market's directional judgment on earnings trends. The "year-to-date price change" is measured by the year-to-date return of the Shenwan primary industry index, ensuring a comparison within the same time window. It is important to note that a positive difference does not automatically equate to a "buy" signal. We further break this down: if the earnings upgrade is positive and the price change is negative, it constitutes the most classic "wrongful sell-off." If the earnings upgrade is positive and significantly exceeds the price increase, even if the price has risen, it still represents a "relative undervaluation" mismatch. Conversely, sectors with a negative difference may face the risk of "performance downgrades but prices have not fallen," or a "Davis Double-Kill" where both performance and stock prices decline simultaneously, requiring differentiated treatment.

Non-ferrous Metals represent the most prominent direction based on this framework. The consensus forecast for net profit attributable to the parent company for the industry in 2026 was revised up from 256.5 billion yuan at the start of the year to 381.9 billion yuan (+48.86%), and the 2027 forecast further increased to 449.9 billion yuan (+50.66%). This results in a two-year average upgrade of 49.8%. However, the sector's year-to-date price change is -0.2%, making the difference the largest across the entire market. Since non-ferrous metals are a strongly macro-driven sector, their intrinsic value can be approximated as a function of the interest rate path, rather than being solely determined by supply, demand, and earnings. Whether the current valuation of the sector can be repaired largely depends on the subsequent interest rate path of the Federal Reserve. The medium-to-long-term underlying narratives that the market has focused on for the non-ferrous metals sector—whether it's supply constraints for related metals, de-dollarization, continued central bank gold purchases, or new demand from AI and grid investments—remain ongoing. The recent decline primarily stems from market concerns following the US-Iran-Israel conflict regarding US inflation control, US economic resilience, and whether the rate-cutting cycle is definitively over. These concerns pressure the sector's valuation within the macro liquidity narrative. Since August, the market has started to price in a marginally stronger expectation of rate cuts, corresponding to US bond yields not continuing to surge but declining slightly. Furthermore, analysts have been continuously upgrading the sector's earnings forecasts, indicating substantive improvement on the earnings side. However, the stock price has only reflected a minor valuation repair. Under the DCF framework, the stock price is determined by both the earnings/cash flow component and the discount rate component. Non-ferrous metals are currently in a state where the earnings side is significantly improving but the discount rate side is under pressure from interest rates. Crucially, the magnitude of improvement on the earnings side far exceeds the compression on the discount rate side. Historically, when the market had higher expectations for Fed rate hikes, it would simultaneously suppress the valuation multiple and the earnings forecast itself. We believe that the magnitude of the earnings upgrade for the non-ferrous metals sector is sufficient that even subsequent rate hikes would still leave it in a state of "wrongful sell-off." Since September 2024, the average real interest rate on long-term US Treasury bonds has cumulatively increased by about 100 basis points. Using the DCF valuation framework P = E / (r - g), where the discount rate r is composed of the US real interest rate plus the sector's risk premium, and assuming a risk premium of 5% and a long-term perpetual growth rate g of 5%, the current denominator r - g is estimated at 3%. In an extreme scenario, if the Fed were to raise rates by an additional 100bp, r - g would expand to 4%. With earnings E holding constant, the sector's valuation would be passively compressed by 25%. In contrast, on the earnings side, the two-year average consensus upgrade is 49.8%, indicating a nearly 50% increase in the earnings base. Comparing the extreme rate hike scenario's 25% valuation compression with the 49.8% earnings upgrade, the improvement in earnings significantly covers the impact from the denominator, showing a clear characteristic of a wrongful sell-off at the stock price level. Looking at the internal structure of non-ferrous metals, we believe US debt issues and the long-term credit concerns surrounding the US dollar support valuations for precious metals. Supply constraints for industrial metals, coupled with structural demand expansion from re-industrialization and new energy, support their demand.

Non-bank Financials have seen their stock price fall 15.6% year-to-date, while the average earnings upgrade for 2026E/2027E is only -2.5%, meaning expectations have barely changed. The difference between the upgrade magnitude and the price change is 13.1 percentage points. The current pessimistic logic for non-bank financials often holds that as trading volumes decline, earnings for the sector will fall, implying that the market's current pricing reflects a reasonable valuation for the peak of the cycle. This view overlooks two variables. First, after a period of adjustment, margin trading balances have risen again to a non-trivial absolute level, providing relatively stable interest income. Second, we are optimistic about a recovery in the A-share market in the second half of the year, which could improve investment returns for the insurance sector's investment side.

Observing mismatches at the secondary industry level, National Defense & Military has seen its stock price fall 15.9% year-to-date, while the average earnings upgrade for 2026E and 2027E is +2.7%, resulting in a difference of 18.6 percentage points. The reasons for the military sector being wrongly sold off may be related to a recent easing of geopolitical risks and, on the other hand, a "liquidity drainage effect" from the high crowding in AI hardware during the first half of the year. However, the fundamental logic of the military sector is not perfectly synchronized with geopolitical risks. Fundamental changes, such as the upgrade of military electronics and information technology and the expansion of arms exports, are ongoing. With current valuations at low levels and an improvement in ROE, we believe the sector exhibits characteristics of "wrongly sold-off assets" and has the basis for a recovery. Since the granularity of primary industry classifications is limited, we have refined our analysis down to the Shenwan secondary industry level. We construct a difference by comparing the "change in consensus forecast net profit attributable to the parent company" (on a FY1 rolling basis) from the current period to the start of the second quarter with the "price change." The ranking of these differences at the secondary industry level reveals internal differentiation within primary industries. Furthermore, as the difference is a percentage-to-percentage comparison of "change in expectations minus price change," it is relatively comparable across industries, unlike absolute amounts. Data shows that for industries like photovoltaic equipment and semiconductors, the difference between expected change and price change exceeds 100%. Additionally, certain sub-sectors within chemicals (agricultural chemicals), military (aerospace equipment II), home appliances (black appliances), food & beverage (snack foods), and media (gaming II) also demonstrate strong configuration cost-effectiveness at the secondary industry level, supplementing the primary industry analysis. It is worth noting that, whether at the secondary or primary industry level, electronics-related sectors are almost "in a league of their own," with expected year-on-year growth rates far exceeding the increase in stock prices. However, the contradiction for electronics lies in its high valuation levels, a situation also seen with telecommunications, another major market theme, before June of this year. Structurally, the various sub-directions within the electronics secondary industry are quite differentiated. Therefore, we believe that the main contradiction for hard-tech sectors like electronics and telecommunications is not short-term business cycles. Instead, they are more significantly influenced in the short term by trading dynamics, the AI CapEx narrative, and potential second-order derivative effects on key price increase segments. Considering this, we introduce a framework for identifying mispricing based on quality factors and valuation: observing the divergence between PB and ROE.

Where to Start

We build a framework using a scatter plot with the PB percentile since 2012 on the horizontal axis and the ROE percentile since 2012 on the vertical axis to study the alignment between valuation and profitability across primary industries. For sectors where the ROE percentile is high but the PB percentile has not fully reflected this (the upper-left quadrant), the business cycle framework still holds allocation value. However, when both ROE and PB enter historically high territory (the far end of the upper-right quadrant), the win rate and risk-reward ratio of the business cycle framework may simultaneously decline. In such cases, introducing valuation protection and a mechanism for confirming business cycle turning points is necessary to correct the original framework. The chart uses long-term percentiles rather than absolute values for both dimensions, eliminating the comparability issue caused by different absolute levels of profitability and valuation across industries. The percentile measures the current indicator's relative position over the past decade or more; a higher percentile means the current valuation or profitability is closer to the historical upper bound. The starting point of 2012 covers multiple complete economic cycles and A-share bull/bear cycles. ROE comprehensively reflects profitability, asset efficiency, and leverage levels, making it the most stable indicator for measuring a sector's business cycle quality, and it is less susceptible to base effects compared to single-quarter net profit growth rates. We choose PB as the valuation proxy variable because PB is sensitive to asset replacement cycles and effectively captures valuation swings in cyclical and growth industries. Using the 50% percentile line as a boundary, the coordinate system is divided into four quadrants, each corresponding to a different business cycle and valuation status. The upper-left quadrant indicates high ROE but low PB. This suggests that business cycle strength has materialized in earnings but valuations have not caught up. It represents a left-side opportunity zone with the highest win rate. The upper-right quadrant indicates high ROE and high PB. This suggests a resonance between business cycle strength and valuation, a zone of trend reinforcement, but one must be alert to the potential decay of this momentum. The lower-right quadrant indicates low ROE but high PB. Valuations may have already priced in overly optimistic future business cycle expectations, and if these expectations fail to materialize, the sector is at risk of a valuation downgrade. The lower-left quadrant indicates low ROE and low PB. Both business cycle strength and valuation are low, potentially requiring a wait for a business cycle turning point.

Why Only 10 ASX 200 Shares?

The PB-ROE framework, together with the previously mentioned earnings upgrade and price divergence framework, jointly confirms the current significant undervaluation of the Non-bank Financial sector. The current ROE percentile for the non-bank financial sector is at the 84th percentile since 2012, a level comparable to high-growth sectors like telecommunications, electronics, and military. However, its PB percentile is below the 10th percentile since 2012. If we focus only on the valuation and ROE situation since the "924" policy pivot, the current valuation percentile for the non-bank financial sector is in the 3rd percentile range since the start of that rally, making it the "cheapest" direction among the 30 primary industries. Meanwhile, its ROE has risen from 6.25% to 11.37%, indicating a potential for future valuation repair.